Ayush Dev

The extension is live now on edge store now! Helps to easily detect and mask custom defined PII information while talking llm. > All rules are 100% local to browser. > No data leak worries even while defining api keys, cards numbers. Ext will detect is ASAP. > Will soon add sync feature. Untill then easy to export/import rules. Also, open sourced the code to keep adding other llm sources. Link:

This is all impressive and all but I am sorry since when this became a real metric? The total assets under management of my own open source projects users is > $100 trillion because it's used by the US government lmao

did we do this right?

Joined Razorpay as Principal Engineer II :) From being a long-time customer to now building parts of the system - it's a full circle. Fintech is a new territory for me - time to get under the hood of how money actually moves. New domain, same guarantees - availability, correctness, performance - just with real money on the line.

The Codex team are hardcore builders and it really comes through in what they create. No surprise all the hardcore builders I know have switched to Codex. Usage of Codex is growing very fast:

Fun Fact: Stripe does logging differently. They take a super interesting approach to log processing, something that helps them debug production issues faster, cheaper, and with far less complexity. Logging data belonging to one request is typically dumped across multiple lines. Stitching together these fragmented logs became a pain as they scaled. So what they do is emit a single, comprehensive log line at the end of every request, and they call it the Canonical Log Line. I just published a video dissecting their entire process and the exact steps they took to implement canonical log lines without breaking their existing log processing systems. Give it a watch. It is a pretty short video, but interesting and fun. Something you can implement at your workplace right away.

This is v0.1. Open, hackable, and already doing things Alexa can't — local LLM inference, real-time sound classification, full developer access. Every product starts somewhere. This is ours.

Quick question: Query time -> 2 seconds. After indexing -> 50ms. Same data. So what was the database doing for 1.95 seconds?